RESEARCH POPULATION GENETICS fromtheanalysisdatasetsindividualscoveredby <10,000 SNPs, with evidence of contamination, or first-degree relatives of others (table S1). We analyzed the data with principal components The genomic history of the Iberian analysis (PCA) (Fig. 1, C and D), f-statistics (11), and qpAdm (12) and summarize the results in Peninsula over the past 8000 years Fig. 1E. We confirmed the robustness of key findings by repeating analyses after remov- Iñigo Olalde1*, Swapan Mallick1,2,3, Nick Patterson2, Nadin Rohland1, ing SNPs in CpG dinucleotides (table S5) that Vanessa Villalba-Mouco4,5, Marina Silva6, Katharina Dulias6, Ceiridwen J. Edwards6, are susceptible to cytosine-to-thymine errors Francesca Gandini6, Maria Pala6, Pedro Soares7, Manuel Ferrando-Bernal8, even in UDG-treated libraries (1). Nicole Adamski1,3, Nasreen Broomandkhoshbacht1,3, Olivia Cheronet9, Previous knowledge of the genetic structure Brendan J. Culleton10, Daniel Fernandes9,11, Ann Marie Lawson1,3, of Mesolithic Iberia comes from three individ- 1,2,3 1,3 1,3 1 2 Matthew Mah , Jonas Oppenheimer , Kristin Stewardson , Zhao Zhang , uals from the northwest: LaBraña1 ( ), Canes1 5 5 Juan Manuel Jiménez Arenas12,13,14, Isidro Jorge Toro Moyano15, ( ), and Chan ( ). We add LaBraña2, who was a Domingo C. Salazar-García16, Pere Castanyer17, Marta Santos17, Joaquim Tremoleda17, brother of the previously reported LaBraña1 (figs. S1 and S2 and table S6), as well as Cueva Marina Lozano18,19, Pablo García Borja20, Javier Fernández-Eraso21, de la Carigüela (fig. S10), Cingle del Mas Nou, José Antonio Mujika-Alustiza21, Cecilio Barroso22, Francisco J. Bermúdez22, and Cueva de la Cocina from the southeast. In Enrique Viguera Mínguez23, Josep Burch24, Neus Coromina24, David Vivó24, northwest Iberia, we document a previously un- Artur Cebrià25, Josep Maria Fullola25, Oreto García-Puchol26, Juan Ignacio Morales25, appreciated ancestry shift before the arrival of F. Xavier Oms25, Tona Majó27, Josep Maria Vergès18,19, Antònia Díaz-Carvajal28, Downloaded from 28 25 29,30,31 farming (Fig. 2A, fig. S5, and table S7). The oldest Imma Ollich-Castanyer , F. Javier López-Cachero , Ana Maria Silva , individual Chan was similar to the ~19,000-year-old 32 33 32 Carmen Alonso-Fernández , Germán Delibes de Castro , Javier Jiménez Echevarría , El Mirón, whereas the La Braña brothers from 34,35 36 Adolfo Moreno-Márquez , Guillermo Pascual Berlanga†, Pablo Ramos-García , ~1300 years later were closer to central European 34 34 37 José Ramos-Muñoz , Eduardo Vijande Vila , Gustau Aguilella Arzo , hunter-gatherers like the Hungarian KO1, with Ángel Esparza Arroyo38, Katina T. Lillios39, Jennifer Mack40, Javier Velasco-Vázquez41, an even more extreme shift ~700 years later in 42 43,44 45 Anna Waterman , Luis Benítez de Lugo Enrich , María Benito Sánchez , Canes1. This likely reflects gene flow affecting http://science.sciencemag.org/ Bibiana Agustí46,47, Ferran Codina47, Gabriel de Prado47, Almudena Estalrrich48, northwest Iberia but not the southeast, where Álvaro Fernández Flores49, Clive Finlayson50,51,52,53, Geraldine Finlayson50,52,53, individuals remained close to El Mirón (Fig. 2A). Stewart Finlayson50,54, Francisco Giles-Guzmán50, Antonio Rosas55, More data from the Mesolithic period, especially Virginia Barciela González56,57, Gabriel García Atiénzar56,57, Mauro S. Hernández Pérez56,57, from currently unsampled areas, would provide Armando Llanos58, Yolanda Carrión Marco59, Isabel Collado Beneyto60, additional insight into the geographical impact David López-Serrano61, Mario Sanz Tormo†, António C. Valera62, Concepción Blasco43, and archaeological correlates of this ancestry shift. Corina Liesau43, Patricia Ríos43, Joan Daura25, María Jesús de Pedro Michó63, For the Neolithic and Copper Age, we model Agustín A. Diez-Castillo64, Raúl Flores Fernández†, Joan Francès Farré65, populationsasmixturesofgroupsrelatedto Rafael Garrido-Pena43, Victor S. Gonçalves30, Elisa Guerra-Doce33, Anatolian Neolithic, El Mirón, and KO1 (Fig. 2A 66 67 68 and table S8). We replicate previous findings of Ana Mercedes Herrero-Corral , Joaquim Juan-Cabanilles , Daniel López-Reyes , the arrival of Anatolian Neolithic–associated an- on March 14, 2019 Sarah B. McClure69, Marta Merino Pérez70, Arturo Oliver Foix37, cestry in multiple regions of Iberia in the Early Montserrat Sanz Borràs25, Ana Catarina Sousa30, Julio Manuel Vidal Encinas71, Neolithic (7, 8, 12); however, sampling from this Douglas J. Kennett10,69, Martin B. Richards6, Kurt Werner Alt72,73, 4,74 9 8 1,2,3 period remains limited and studies of larger sam- Wolfgang Haak , Ron Pinhasi , Carles Lalueza-Fox *, David Reich * ple sizes and additional sites will be important to shed further light on the interaction between We assembled genome-wide data from 271 ancient Iberians, of whom 176 are from the the incoming farmers and indigenous hunter- largely unsampled period after 2000 BCE, thereby providing a high-resolution time transect gatherers. For the Middle Neolithic and Copper of the Iberian Peninsula. We document high genetic substructure between northwestern Age, we reproduce previous reports of an in- and southeastern hunter-gatherers before the spread of farming. We reveal sporadic crease of hunter-gatherer–related ancestry after contacts between Iberia and North Africa by ~2500 BCE and, by ~2000 BCE, the 4000 BCE (6, 7, 12, 13), with higher proportions ’ replacement of 40% of Iberia s ancestry and nearly 100% of its Y-chromosomes by people in groups from the north and center. Using our with Steppe ancestry. We show that, in the Iron Age, Steppe ancestry had spread not only observations about population substructure in – – into Indo-European speaking regions but also into non-Indo-European speaking ones, and theMesolithicasareferenceframe,weshowthat we reveal that present-day Basques are best described as a typical Iron Age population the hunter-gatherer–related ancestry during those without the admixture events that later affected the rest of Iberia. Additionally, we document periods was more closely related to later north- how, beginning at least in the Roman period, the ancestry of the peninsula was transformed western (Canes1-like) hunter-gatherers than to by gene flow from North Africa and the eastern Mediterranean. the El Mirón–like hunter-gatherers (Fig. 2A), pro- viding clues about the source of this ancestry. he Iberian Peninsula, lying at the extreme single-nucleotide polymorphisms (SNPs) (2, 3) Our Copper Age dataset includes a newly re- southwestern corner of Europe, provides to generate genome-wide data from 4 Mesolithic, ported male (I4246) from Camino de las Yeseras an excellent context in which to assess the 44 Neolithic, 47 Copper Age, 53 Bronze Age, (14) in central Iberia, radiocarbon dated to T final impact of population movements en- 24 Iron Age, and 99 historical-period Iberians 2473–2030 calibrated years BCE, who clusters tering the continent from the east as well (Fig.1,AandB,andtablesS1andS2).Wealso with modern and ancient North Africans in the as interactions with North Africa. To study the generated 26 radiocarbon dates (table S3). We PCA (Fig. 1C and fig. S3) and, like ~3000 BCE genetic impact of prehistoric and historic events co-analyzed the new genomic data with previ- Moroccans (8), can be well modeled as having in Iberia, we prepared next-generation sequenc- ously reported data from 1107 ancient individ- ancestry from both Late Pleistocene North Africans ing libraries treated with uracil-DNA glycosylase uals, including 132 from Iberia (Fig. 1B) (2, 4–9), (15)andEarlyNeolithicEuropeans(tablesS9and (UDG) (1) and enriched them for ~1.2 million and 2862 present-day individuals (10). We filtered S10). His genome-wide ancestry and uniparental Olalde et al., Science 363, 1230–1234 (2019) 15 March 2019 1of5 RESEARCH | REPORT markers (tables S1 and S4) are unique among cestry on the X-chromosome (table S14 and fig. in Western Europe but overlap genetically with Copper Age Iberians, including individuals from S7), a paradigm that can be exemplified by a Iron Age populations (Fig. 1D) showing substan- sites with many analyzed individuals such as Bronze Age tomb from Castillejo del Bonete tial levels of Steppe ancestry. Sima del Ángel, and point to a North African containing a male with Steppe ancestry and a In the historical period, our transect begins origin. Our genetic evidence of sporadic contacts femalewithancestrysimilartoCopperAge with 24 individuals from the 5th century BCE with North Africa during the Copper Age fits Iberians. Although ancient DNA can document to the 6th century CE from the Greek colony of with the presence of African ivory at Iberian that sex-biased admixture occurred, archaeolog- Empúries in the northeast (19)whofallinto sites (16) and is further supported by a Bronze ical and anthropological research will be needed two main ancestry groups (Fig. 1, C and D, and Age individual (I7162) from Loma del Puerco to understand the processes that generated it. fig. S8): one similar to Bronze Age individuals in southern Iberia who had 25% ancestry re- For the Iron Age, we document a consistent from the Aegean, and the other similar to Iron lated to individuals like I4246 (Fig. 1D and trend of increased ancestry related to Northern AgeIberianssuchasthosefromthenearbynon- table S16). However, these early movements and Central European populations with respect Greek site of Ullastret, confirming historical from North Africa had a limited impact on to the preceding Bronze Age (Figs. 1, C and